Best AI for Risk management and position sizing
Calculate proper position size using fixed fractional, volatility-adjusted, or Kelly criterion methods — and run Monte Carlo simulations to see expected drawdown across different sizing approaches.
ChatGPT
ChatGPT with Advanced Data Analysis (Python code interpreter) is the strongest tool for position sizing because the right answer often requires running scenarios, not just one calculation. ChatGPT can execute Python to run Monte Carlo simulations across 1000+ trades using your win rate and average R-multiples, calculate Kelly criterion with the half-Kelly safety adjustment most pros actually use, and visualize how different risk-per-trade settings change your expected drawdown distribution. Plus and Pro tiers ($20/month and $200/month) include code execution; the free tier does not. Use ChatGPT here specifically — not Claude — because the code interpreter is what makes the scenario work happen.
Open ChatGPTRun a complete position sizing analysis for the trade below: 1. Fixed fractional sizing at 0.5%, 1%, and 2% risk 2. Volatility-adjusted size using 14-day ATR 3. Kelly criterion size and half-Kelly (with my historical win rate and average R) 4. Monte Carlo simulation: expected max drawdown over 100 trades at each sizing level 5. Recommendation with reasoning Account: $[X] | Win rate: [Y%] | Avg winner: [Z]R | Avg loser: [W]R | Entry: $[A] | Stop: $[B]
See the difference
Before vs. after using this prompt
Trader has a $50,000 account and likes a setup in NVDA at $880 with a stop at $860. They eyeball it and buy 50 shares — 'feels about right.' They've actually risked $1,000 (2% of account) on a single trade. After three losses in a week they're down 6% and emotionally compromised.
Same trader pastes the setup into ChatGPT with their stats. The Monte Carlo simulation shows 2% risk has a 41% chance of a 15%+ drawdown over 100 trades vs. 11% chance at 1% risk. They size 25 shares ($500 risk, 1% of account). Three losses later they're down 3% — uncomfortable but functional, with another 30 trades of edge ahead of them.
Claude
Better when you want clear conceptual explanation rather than executed simulations. Claude is excellent at walking through why Kelly criterion can be dangerous in practice or how volatility-adjusted sizing differs from fixed fractional, but won't run the actual Python by default in the consumer interface. Use Claude to understand the framework; use ChatGPT to run the numbers on your specific account.
Open ClaudeFrequently asked
Should I actually use Kelly criterion in real trading?
Full Kelly is mathematically optimal but practically dangerous because it assumes you know your true win rate and edge precisely — which you don't. Most professionals use half-Kelly or quarter-Kelly to account for estimation error. For a retail trader with 50–200 trades of history, flat 0.5–1% fixed fractional sizing is safer than any Kelly variant because your stats aren't reliable yet.
What's volatility-adjusted position sizing and when is it worth using?
Volatility-adjusted sizing uses Average True Range (ATR) to make all trades risk the same dollar amount regardless of how wide the stop has to be. Worth using when you trade across instruments with very different volatility — sizing AAPL the same as TSLA without adjusting gives you wildly different real risk. Less critical if you only trade large-caps with similar volatility.
How much should I actually risk per trade?
For most retail traders, 0.5–1% per trade is the right range. Above 2% means a 10-trade losing streak takes you down 20% — survivable but psychologically devastating. Below 0.25% means your edge can't compound meaningfully. The right number is the one where you can take a loss without it changing how you trade tomorrow.